A VVT (Variable Valve Timing) apparatus has conventionally been known that changes the timing at which an intake valve or an exhaust valve is opened/closed, that is, the opening/closing phase (crank angle) according to an operating condition. Generally, in the variable valve timing apparatus, the phase is changed by rotating a camshaft, which opens/closes the intake valve or exhaust valve, relative to a sprocket or the like. The camshaft is rotated by an actuator such as a hydraulic or electric motor.
In order to accurately control the valve opening/closing phase (valve timing) using such a variable valve timing apparatus, it is necessary to prevent error in detecting the actual phase of valve opening/closing. In order to reduce the detection error, it has been a common practice to set the valve opening/closing phase at a prescribed reference position that is limited mechanically, and to learn the error in the detected value of valve opening/closing phase at that time as an offset.
In an intake valve driving apparatus disclosed in Patent Document 1 (Japanese Patent Laying-Open No. 2004-340013), target working angle and target phase are set by adding a learning correction value or values, whereby variation in variable valve control is corrected. Particularly, according to Patent Document 1, the effect of suppressing variation is improved when the learning operation for updating the learning correction value is done on a low-speed, low-load side.
In a variable valve timing apparatus disclosed in Patent Document 2 (Japanese Patent Laying-Open No. 2004-156461), reference position of the valve timing is learned under prescribed learning conditions (for example, every time engine operation starts), to ensure detection accuracy of the actual valve timing. Further, according to the disclosure, when learning is not complete, it is determined that the detection accuracy is low, and the rate of change in valve timing is limited. Consequently, damage to the apparatus caused by a movable portion hitting a stopper or the like at high speed can be prevented.
As one type of variable valve timing apparatus, a mechanism has been used in which, when an actuator operating a movable portion for changing the valve timing is stopped, the movable portion is urged by a spring or the like, or operation of the movable portion is limited by a lock-pin or the like, so that the valve timing is automatically returned to the reference position. In such a mechanism, the reference position learning is naturally done at the time of such return.
In a variable valve timing apparatus having such a mechanism that the valve timing is changed by an amount in accordance with the operation amount of the actuator and the valve timing is fixed when the actuator is stopped, it is necessary to execute the reference position learning for ensuring accuracy in detecting actual valve timing, in consideration of protection of apparatuses as well as operation energy (power consumption) of the actuator. Patent Documents 1 and 2 mentioned above do not describe any specific contents of reference position learning from such a viewpoint.